Python 64 bit!

I got python 64 bit to compile, but it required a number of tricky steps.

First, this guy has the instructions I followed:
captnswing

However, it didn't work entirely as advertised. I ran the configure as advertised:

./configure --enable-framework=/Library/Frameworks \
--enable-universalsdk=/ \
MACOSX_DEPLOYMENT_TARGET=10.5 \
--with-universal-archs=all \
--with-readline-dir=/usr/local


then the make install, but /usr/local/bin/python pointed to the wrong place, so I replaced the symbolic link in my python path with the correct one:

sudo rm /Library/Frameworks/Python.framework/Versions/2.6/bin/python
sudo ln -s /Library/Frameworks/Python.framework/Versions/2.6/bin/python-64 /Library/Frameworks/Python.framework/Versions/2.6/bin/python


Now python is 64 bit:

eta ~$ python -c "import sys; print sys.maxint"
9223372036854775807


I haven't checked whether it works yet though...

Update: Had to reinstall with gnu readline installed. Also have to install PyQt4 and might have to recompile numpy...

numpy won't compile with python 2.6.2:
C compiler: gcc -arch i386 -arch ppc -arch ppc64 -arch x86_64 -isysroot / -fno-strict-aliasing -fno-common -dynamic -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes

compile options: '-Inumpy/core/include -Ibuild/src.macosx-10.5-universal-2.6/numpy/core/include/numpy -Inumpy/core/src -Inumpy/core/include -I/Library/Frameworks/Python.framework/Versions/2.6/include/python2.6 -c'
gcc: build/src.macosx-10.5-universal-2.6/numpy/core/src/_sortmodule.c
In file included from numpy/core/include/numpy/ndarrayobject.h:26,
from numpy/core/include/numpy/noprefix.h:7,
from numpy/core/src/_sortmodule.c.src:29:
numpy/core/include/numpy/npy_endian.h:33:10: error: #error Unknown CPU: can not set endianness
lipo: can't figure out the architecture type of: /var/folders/ni/ni+DtdqFGMeSMH13AvkNkU+++TI/-Tmp-//cceaWIvZ.out
In file included from numpy/core/include/numpy/ndarrayobject.h:26,
from numpy/core/include/numpy/noprefix.h:7,
from numpy/core/src/_sortmodule.c.src:29:
numpy/core/include/numpy/npy_endian.h:33:10: error: #error Unknown CPU: can not set endianness
lipo: can't figure out the architecture type of: /var/folders/ni/ni+DtdqFGMeSMH13AvkNkU+++TI/-Tmp-//cceaWIvZ.out
error: Command "gcc -arch i386 -arch ppc -arch ppc64 -arch x86_64 -isysroot / -fno-strict-aliasing -fno-common -dynamic -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -Inumpy/core/include -Ibuild/src.macosx-10.5-universal-2.6/numpy/core/include/numpy -Inumpy/core/src -Inumpy/core/include -I/Library/Frameworks/Python.framework/Versions/2.6/include/python2.6 -c build/src.macosx-10.5-universal-2.6/numpy/core/src/_sortmodule.c -o build/temp.macosx-10.5-universal-2.6/build/src.macosx-10.5-universal-2.6/numpy/core/src/_sortmodule.o" failed with exit status 1


That sucks.

Hosting on my mac

I'm hosting my website off my mac; it should be accessible to the outside world now: eta.colorado.edu.

A few things went into this....

1. Don't install the fink version. The two versions clash and depending on how you access your computer you could end up looking at entirely different directories (e.g. localhost and eta.colorado.edu referred to different sites for a while)
   
2. The configuration file is /private/etc/apache2/httpd.conf
   
3. I needed to change DirectoryIndex to index.htm (from index.html)
   
4. uncommmented "LoadModule php5_module libexec/apache2/libphp5.so"
   
5. had to allow override so that .htaccess files would work.
   

Mac stuff cont'd

Trying to get Apache server to run, and it's just a pain.

I frequently forget how to update the locate database because it's different on macs. Marcos' Mac Singularity has the instructions .

In short:
sudo /usr/libexec/locate.updatedb

Google has failed me

So, either no one on the internet has successfully installed 64-bit python on a mac (which I find nothing short of impossible) or google has totally failed. None of the hits for any combination of 64 bit, python, mac, etc. has shown a useful result. Wow.

Caffeine

Caffeine consumption stopped around mid-October. I think it was when I got back from observing so October 20 give or take. Re-started consumption mid-March. Date is hard to say... I think it was Sunday before Spring Break, but hard to say because I still took it easy until ~last week.

Conclusions? I'm probably more tired in the mornings with caffeine than without. Also, probably perform better @ ulty when I'm not addicted. However.... not really conclusive. Just addictive.

weekend

What did I do this weekend?
-upgraded PyRAF from 1.6.1 to 1.7
-switched matplotlib backends from TkAgg to Qt4
-put up a new article on molecularclouds


So... totally unproductive, but only in an awesome way?

I tried to switch PyRAF from X11 to Aqua a la Ticket 86 on the PyRAF trac site, but it failed with a can't-load error.

Also, updated readcol documentation so that it makes sense. Sam has used it, Robbie might.

April 1st, 2009

No April fools' day would be complete without scientists' contributions:

http://arxiv.org/abs/0903.5308
http://arxiv.org/abs/0903.5321
http://arxiv.org/abs/0903.5377

although it's not obvious whether .5308 is actually intended to be humorous

Astronomy gets some recognition

First off, this guy does some amazing work, like measuring the galactic rotation curve. He got recognized by the Colbert Report. Sweet.

My works featured...

Curiously, one of my wikimedia commons images has been included in this article: http://www.scienceinschool.org/2008/issue10/tamaradavis, though the image has little to do with the subject in question. Noodle (can't remember her real name... that's sad on my part) is in the picture. Neat!

This might not be something to be proud of:
http://www.denverptc.org/denver.html

GC was featured in:
http://www.strudel.org.uk/blog/astro/000855.shtml

ps2pdf keep resolution, crop to bounding box

ps2pdf -dEPSCrop -dAutoRotatePages=/None -dAutoFilterColorImages=false -dColorImageFilter=/FlateEncode -dUseFlateCompression=true l007_displaycrop.ps
ref: this post

Making a postscript plot of a gigantic fits image


map = readfits('MOSAIC.fits',hdr)

crpix1 = sxpar(hdr,'CRPIX1')
crpix2 = sxpar(hdr,'CRPIX2')
crval1 = sxpar(hdr,'CRVAL1')
crval2 = sxpar(hdr,'CRVAL2')
cd1_1 = sxpar(hdr,'CD1_1')
cd2_2 = sxpar(hdr,'CD2_2')

x = lindgen(n_e(map[*,0]))
y = lindgen(n_e(map[0,*]))
l = (x-crpix1)*cd1_1+crval1-360
b = (y-crpix2)*cd2_2+crval2

imdisp,map,/axis,xrange=[max(l),min(l)],yrange=[min(b),max(b)],range=[-1,8]


This code makes use of imdisp.pro.

MOSAIC data reduction

MOSAIC reduction is very difficult.
http://www.noao.edu/noao/noaodeep/ReductionOpt/frames.html has the official instructions.

Important things:
Have the latest version of MSCRED and MSCDB installed. Both will give cryptic errors or "cannot open file" errors (because the files don't exist) otherwise.



mscred
setinstrument kpno CCDMosaThin1
msccmatch obj09*.fits coords="!mscgetcat $I $C" search=60 rsearch=1 nfit=30 accept=yes interactive=no fit=no


Site visits, comps

Amusing - visits to my site increased steadily leading up to Jan 9, then crashed precipitously.

LaTeX: replace double quotes with tex quotes

People often make the mistake of putting " in place of `` in LaTeX documents. To repair this, the only easy solution is something like:
s/\(\s\)"/\1``/g
in VIM or sed.

I've seen alternate solutions posted but they're all too complicated. One recommendation was to switch to XeTeX, which sounds like overkill, and others all stated that a complicated perl script is required. The latter is technically true, but why isn't such a thing readily available?

Any other ideas?

Skim: a PDF reader that allows annotations

from Sam: http://skim-app.sourceforge.net/

neat

skill set

is it a marketable skill that I can bring down almost any computer system I'm given access to entirely unintentionally?

post rate

my post rate is too low I need to up it so that my front page doesn't show the same posts any more

scary

I think this is my worst nightmare:
http://antwrp.gsfc.nasa.gov/apod/ap081203.html

LA. Not the smiley.

.inputrc

New unix stuff!

This is, again, stuff I'd searched for forever but never found the right collection of terms. This is what bash readline is REALLY all about.

if you make an inputrc file, you can get awesome readline capabilities (similar to ipython). This post at lifehacker showed me how to do it. This site has a more complete description of how to use readline/inputrc.

Even better: VI command line editing .

There's also something called magic space but I don't know what purpose it serves all the time... it will replace 'magic' characters like !* and !& and !! with whatever command they represent if you hit space immediately after typing them (explained better here).

Wow. Awesome.

MOVIES!

Making movies is surprisingly difficult. No matter what language you use, apparently true movie files can only be made by stringing together images, i.e. there is no native movie-producing feature. Gnuplot can do some simple animations but to do anything sophisticated you need to start delving into variables, and for that I switch to a real language.

So, I returned to python. As usual, it took no more than a few hours of coding and learning to come up with something. But it bothers me that it took that long: I still think python is most deficient in its failure to create a default column-text reader like 'readcol' in IDL. I can't complain that much, though: I wrote my own in about 5 minutes.

Anyway, the key is to use the .set_xdata and .set_ydata functions of a plot to update a canvas. I still don't have nearly as high a plotting speed as I'd like, but it works alright if I don't display to screen. Probably a different backend would be more effective but I don't like to mess with backends.

I use savefig(filename,dpi=50) to reduce the image quality so that it's easier for the animator to handle.

ImageMagick's convert can be used to stitch any kind of image into a movie given that you've installed an mpeg2 encoder (fink gave me mpeg2vidcodec ). The command is very simple:

convert -size 300x300 *.png movie.mpg


I had to use a smaller image size because a series of 1000x12kb files somehow chomped ~6-8 GB of RAM and swap space.